James Cassell's Blog

Sunday, November 16, 2008

Corrupted Vista Registry


About a two months ago, I had a problem with Windows. I got the following error: "Windows could not start because the following is missing or corrupt: \Windows\System32\config\SYSTEM" I went to the Help Desk at the VCC, and surprise, surprise, after several hours, they weren't able to fix the problem.

About a week later, solved the problem myself, and am documenting it for you here.


So, the solution is to boot into a recovery environment, and to copy the file at "C:\Windows\System32\config\RegBack\SYSTEM.OLD" to "C:\Windows\System32\config\SYSTEM", making sure to back up the old "C:\Windows\System32\config\SYSTEM" first, just in case. As far as "recovery environments" go, I just booted into Linux, though a Live CD, or even the Windows installation disc would work just as well.

This same problem happened to me about two weeks ago, and this same procedure saved my system again.

There is a Microsoft KB article that describes how one could solve the problem for Windows XP, but that wasn't much help for the same sort of problem on Vista.

Labels: , , ,

Monday, September 08, 2008

Never Let Windows XP Touch Your Partition Table

The other day, I decided to re-install XP. I have a triple-boot system; on my ThinkPad, I have Vista, XP, and Fedora. I told the XP installer to delete the partition that had my old install of XP, and when I told it to put a new one in its place, it told me that I already had four primary partitions.

My partition table was a follows: first primary partition: Vista; second primary partition: XP; third primary partition: boot partition for Fedora; fourth primary partition: extended partition which holds: 2 encrypted partitions for Fedora.

After the XP installer touched my partition table, the I could only boot into Vista. GParted saw my entire disk as "uninitialized," or basically, empty. At this point, I was in a slight panic; I had a lot of important stuff in my Fedora partitions.

My eventual solution was both tedious and dangerous. I basically edited the partition table by hand, using the command line tool sfdisk. I did this using the Fedora 9 Live CD. This time, I had gparted create an empty NTFS partition, and I told XP to just use that, and I let it format it when it asked, which turned out to be a mistake. This caused it to mess up my partitions again, and I had to use sfdisk to set them straight. I now have a working setup, as I had before re-installing XP.

The moral of this story happens to be the title of this post: Never let Windows XP touch your partition table.

Labels: , , , ,

Saturday, September 06, 2008

Broken Ctrl and Shift Keys

I have been having a very intermittent problem in Linux where my Ctrl and Shift keys would stop working. This prevented me from typing a question mark, as well as preventing me from entering my passwords when they were required. (All but my most insecure passwords require the use of the shift key.) Additionally, this breaks many, many keyboard shortcuts. I had noticed that this problem seemed to show itself whenever I used a program that captured the mouse and keyboard, such as a remote desktop application, or a virtual machine application.

Today, after having failed many times in the past, Google helped me find a solution that didn't require rebooting my machine (which was my only-known solution previously.) In a forum somewhere, someone said that fidgeting about with setxkbmap could sometimes help. It turns out that he was correct. If this happens to you, you can type, "setxkbmap dvorak; setxkbmap us" (without the quotes) into the command line. It worked very well for me, but your mileage may vary.

(Now, if only someone were to make a post like this whenever they solved an obscure computer problem. It would make Google's job much easier.)

Labels: , , ,

Tuesday, September 02, 2008

An Unexpected Adventure

This past Saturday, I woke around noon to the sound of my cell phone. Strangely, it wasn't a telemarketer calling to tell me that the non-existent warranty on my non-existent car was about to expire. One of my friends was having a computer problem, the likes of which I had never heard before. One thing that I have noticed semi-recently is that it is difficult to troubleshoot an unknown computer problem over the phone. (If it's something I've run into before, that's a different story.)

Technical Aspect

Not that it will interest the reader, but I'll describe the symptoms of the problem anyway. (Maybe someone who has a solution will post it in a comment.) Windows Vista would boot fine to the login screen. After typing the password was when the problems started appearing. The "Welcome" message would show for a few seconds, then, when the desktop should have appeared, only a royal blue background appeared. At this point, the mouse was functional (i.e., it moved around,) but there was nothing to click on. It seemed as if explorer had never started. All of the normal approaches to diagnose such a problem (e.g., Ctrl+Alt+Del, Ctrl+Alt+Esc, etc.) were useless. I asked on IRC, but no one had any solutions. Neither did Google. As a side note, it would boot into safe mode.

Eventually, we were able to get the desktop to come up by going into msconfig and disabling all of the startup items as well as all of the (non-essential) services (as defined by msconfig.) Upon re-enabling the services, we were able to get a functional desktop for about one boot, but when we rebooted, the initial problem re-appeared.

I didn't want to try the same solution again because, as Einstein said, it is insane to do the same thing over and over again and expect different results. My next suggestion would have been to get the computer re-imaged. We had been working on this problem for over five hours by this point, and I didn't want to send my friend away having wasted so much time. Since the Help Desk wouldn't be open until the following afternoon, I proposed another solution. Last year, I had taken a backup image of my computer the day that it was issued. I still had this backup, (that consisted of six DVDs,) and proposed that we do a restore from this image. We did so, followed by downloading all of the updates that had been released since. This second option took about six hours. The system was now in a pristine state.

Human Aspect

As can be inferred from the above, this ended up being an all-day event. Some may exclaim, "what a boring way to spend one's Saturday!" I look at it differently. In addition to the fact that "...we know that all things work together for good to them that love God..." (Romans 8:28, KJV), I can't think of a better way to spend a Saturday than getting to better know my friends. In case you know nothing about humans, people don't sit and stare blankly at one another for more than eleven hours, while waiting for a computer to do its thing. Another of my friends joined us about half-way through, and hung out until the end. I met both of these friends at Silver Bay, where, many will agree, the best times since heading off to college, have taken place. This opinion was voiced several times through the day.

While this day may have been a bad day for the victim of the computer problem, it was one of the best days that I have had since returning to Rensselaer. The only thing that I would change about the day is my allowing the conversation to be steered to my planned activity for the day: sleeping into the late afternoon. This caused my friend to feel bad for waking me, and, in turn, made me feel guilty for causing this bad feeling. This could become a vicious circle. Aside from that, I'd say that the day was an excellent unexpected adventure!

Labels: , , ,

Tuesday, August 26, 2008

Small Dorm Room

Last Spring, there was a "lottery," where the order in which students would get to choose their rooms was chosen. I didn't get too bad of a number, but by the time my turn came around, my preferred choices for a room were gone. (These were Davidson, or Nugent Halls.) I ended up with a small single in the Quadrangle, otherwise known as "The Quad." One thing that I declared to be a requirement for my room was air conditioning. The Quad does have air conditioning, which I am happy about.

When I first got into my room, I was really disappointed at how small it really was. There was almost no room to move around. I had planned to mitigate the small space by lofting my bed, and placing my desk underneath. This turned out to not be an option for a couple of reasons. First, my bed has drawers built-in under the mattress, which loses me two feet of vertical space. Second, the ceiling is only eight feet high, and the regulations state that there must be at least three feet of clearance between the top of the mattress and the ceiling. The combination of these two would have left me with only about three feet under the bed, which is insufficient to fit a desk. Another peculiarity of the room is that it is more narrow than it is tall.

When I got into the room, the bed and the desk were parallel, with about 2 feet between. This was not sufficient space to pull my chair out from the desk and comfortably sit in it. What I ended up doing was to re-arrange the furniture in the room. I rotated the bed ninety degrees, which was in and of itself a challenge, as the room wasn't wide enough to properly do so. I pushed the bed as close to the window as I could, which was several feet away because the air conditioner kind of got in the way. I have my servers as well as printer in this space between my bed and the window. Now that I have done all this re-arranging, I have a consolidated floor space, where I can comfortably pull out my chair to sit in.

At this point, I am satisfied with my room, and my initial concerns have been mitigated. I am really enjoying the air conditioning.

Labels: , , , ,

Sunday, February 24, 2008

Hard Drive Encryption Vulnerability

Within the past few days, a vulnerability in the implementation of hard drive encryption was publicized. The problem applies to just about all hard drive encryption schemes -- Windows and Linux alike.

The Problem

The root of the problem is the same as the problem with DRM; the computer must hold the decryption key in order to be able to make use of the encrypted resource. This has been implemented by storing the key in RAM, which, common knowledge tells us, loses its contents when power is lost. This common knowledge assumption has a caveat that is not common knowledge; this loss of memory is not instantaneous. The memory may hold its contents for minutes after it loses its power. If the RAM is chilled before having its power cut, it will hold its contents for hours.

There are several ways that this can be exploited. If someone steps away from their laptop for a moment, an attacker could cut the power, connect an external device from which to boot, and their device could copy the contents of the memory onto itself. Somewhere in this copy of the memory would be the encryption key, which can be used to read the contents of the drive.

A Possible Solution

The article I referenced says that "There seems to be no easy fix for these problems." The hardest part of that assertion to contradict is the "easy" qualifier. However, upon reading about this problem, a seemingly obvious solution came to me. Modern processors have plenty of registers as well as on-board cache. Why not reserve one or more of these registers to hold an encryption key? If using a register would be too expensive, surely it would not be too expensive to use some of the plenteous cache that modern processors possess.

Upon considering this, I realized that people may very well want to have more than one encryption key. A possible solution for this is to store a key (say, a 256-bit key) that is randomly generated each time the system is booted. It could be stored in the manner I described above. This key could then be used to encrypt any other encryption keys before they are stored in RAM, and to decrypt them anytime they are retrieved. The only thing that needs to be done is to ensure that this randomly-generated key does not ever find itself in RAM.

A Word of Caution

Before writing this, I searched through the several hundred comments on the article, and found that several other people had mentioned the processor cache, asking if it were vulnerable. This indicates to me that this solution is, to some extent, common sense. It would be necessary to have security experts examine this approach before anyone tried to implement it, so as to avoid another disaster similar to what happened with the fatal weaknesses of WEP.

Labels: , , ,

Tuesday, January 29, 2008

File Server From Scratch

Back Story

Back in early December before Christmas break, my desktop computer, which had been serving as a server, was running very low on disk space. My immediate solution was to buy another hard drive. I ordered a 750-GB hard drive from NewEgg. As Murphy's Law would predict, my roommate's LCD TV, which I had been using when I needed a screen on my desktop, bit the dust (stopped working). I had no way to tell the BIOS that there was a new disk, and to keep booting from the old disk. I tried to reconfigure the BIOS from memory without any visual feedback, but only managed to make the system unbootable.


Fast forward a few weeks to the middle of vacation. I had, by this point, decided that I was going to build myself a storage server with RAID-5 redundancy. I had several decisions to make, the first and most important of which was whether I should go with Intel or AMD. This would dictate my selection of motherboards, as Intel motherboards and AMD motherboard are mutually exclusive. My second major decision was whether I would go with hardware RAID or software RAID. Obviously, hardware RAID is the better option if money is not a factor, but software RAID cost is very low. In reality, software RAID is not free as it appears to be, but rather, the cost is hidden in the fact that motherboards with many SATA ports are significantly more expensive. As far as the choice of processor manufacturer, there is always a battle going on between Intel and AMD. AMD had been winning until Intel released their Core 2 platform. Since then, AMD has come back with their Phenom processor, which is marginally better than the Core 2, but not enough so to justify the premium price.

While considering the previously mentioned choices, I decided to go ahead and order a case to put everything in. I was initially considering a traditional rack-mount server case, but found them to be prohibitively expensive. I ended up ordering a Cooler Master "Ammo 533" case that was on sale and has subsequently been discontinued. This arrived at my house about a week before I ordered any other parts on-line.

My best friend informed me of a computer show that was going to be taking place at the Dulles Expo Center in Chantilly, Virginia, so I decided that I would look there for computer parts, hopefully at greatly-discounted prices. He and I drove out there to have a look. For the most part, prices were not particularly competitive with on-line retailers. The prices were often within $5 or so, but I think that is worth being able to return an item in the case of failure or DOA. I actually ended up buying a power supply at the show (in addition to some canned air). Later, I found out that I had paid way too much for the power supply and that it was missing some parts that should have come in the box. My friend and I have subsequently decided to boycott the show as the only purpose it serves is to rip people off (or rather to trick them out of their money by offering them sub-par merchandise).


By this time, it was about a week before I had to go back to school, so the urgency of the project went up significantly as I would have little to no time to do it once school was back in session. I finally decided to build an Intel-based machine. I looked briefly on eBay for good prices on the processor as I had done when upgrading my Alienware notebook. No such luck -- it's very hard to compete with NewEgg. I decided to go with the Core 2 Duo: E6750 for the processor as it seemed to be the best balance of price and capability. I went with the Gigabyte GA-P35-DS3P Rev 2.0 motherboard as it had plenty of SATA ports and mostly good reviews. For the memory, I went with the G.Skill 4GB(2 x 2GB) DDR2 800 Dual Channel Kit. I went with the slower memory because the faster stuff was prohibitively expensive. All that was on one order, and I paid for rush processing. I must express my displeasure with rush processing -- it didn't speed up my order at all. They didn't reimburse me the rush-processing charge, on the notion that they shipped the same day I ordered, when in actuality they only got out the information to the carrier the same day (it was around 2300, 11:00 P.M.). The carrier didn't actually get the package until the next afternoon.

If you've ever built a computer, you may have noted that the pieces that I have bought to this point don't form a complete system. I decided to buy neither an optical drive, nor a video card. My reasoning was that a server doesn't really need these things as it is primarily accessed over the network. I also had two optical drives in my older desktop computer, one of which could be transferred to the server. For the time before I returned to school, I was just going to borrow an optical drive and the video card from my mom's (aging) desktop. The day after placing my initial order (the one for the processor, motherboard, and memory), I decided to order a 750-GB hard drive so that I could have an operating system up and running. I now had two of these drives (the second was back in my dorm), which were made by different companies to reduce the chance of simultaneous failure. I just needed a third to have a RAID-5 setup.


My two orders arrived on the same day, and I set out building my first home-built computer. I first put the power supply into the case, which was a no-brainer -- just screw it in. Next, according to the directions that came with the processor, I was supposed to install the processor and heat sink onto the motherboard. I was surprised that the processor didn't have any pins. It had flat contact points, which are called "lands" -- thus, why the socket type is called LGA: Land Grid Array. With this setup, the protruding contact is part of the motherboard. After I had inserted the processor, another decision came up. I had to decide whether to use the thermal interface material that was pre-applied to the stock heat sink, or to use my premium Arctic Silver 5 that I had left over from my when I upgraded my Alienware notebook's processor. After some brief Internet research, I decided to go with the Arctic Silver 5. I used some alcohol to get rid of the stock thermal compound, and applied the Arctic Silver 5. Next, I snapped the heat sink onto the motherboard. This was the most nerve-wracking part of the whole process -- when I pressed down on the clips, the entire motherboard bent terribly from the stress.

The next thing to do was to install the motherboard in the case. There were several stand-offs designed to screw into the case, and the motherboard into them. I came across a minor snag here (for which the internet gave me no solutions). Two of the stand-offs were slightly different from the others -- most of them had a flat top, but two had a pointy top. It turns out that the pointy-topped ones serve to help line up the rest of the stand-offs with the holes in the motherboard -- the pointy ones sink into the holes slightly such that having two of them provides rotational momentum for the whole motherboard. My recommedation for these pointy stand-offs is simply to make sure they are relatively far apart from one another. After getting the motherboard lined up, the task of screwing it in remained. This was rather simple (tightining screws can't get that complicated). After the motherboard was in, I installed the memory, which was as simple as pushing until it clicked. I proceded to connect all the connectors from my case -- such as power and audio among others -- to the motherboard.

Next, came time to connect the power supply to the motherboard. There were two points where I thought I had an unusable power supply; the first was when I realized my power supply had a 20-pin connector and the motherboard had a 24-pin connector. Thankfully, the 20-pin fits, leaving 4 pins without a connection. There was also a separate 4-pin connector that went to the motherboard. The second time I thought I had a bad power supply was when I plugged in the system, flipped the power switch, and nothing happened -- thankfully, it was a silly mistake on my part; I never pushed the regular power button on the front of the case. I installed my hard drive with ease due to the tool-less design of the case. I also temporarily installed the borrowed optical drive and video card, also with ease.


Now came the moment of truth (it was also around 0300 or 3:00 A.M. by this time). I inserted my 64-bit Fedora 8 live DVD, and watched the system boot. I was ecstatic! The system booted perfectly with no glitches, whatsoever! I may sound over-excited here, but with every one of my systems (all notebook computers, granted), it took some labor to get Fedora to work properly. Everything worked seamlessly. I immediately installed Fedora to the hard drive so that I would have a fast-booting system and downloaded the updates. The next thing I did was run SpinRite on the drive to make sure it was in decent shape. Since it takes many hours to do its thing, and because it was getting late, I went to bed at this point.


One thing I really like about my case is its tool-less design. After the initial install, all upgrades (excluding a motherboard upgrade) are tool-less. Expansion cards can be swapped out thanks to a clever latching mechanism, hard drives and optical drives can be swapped out by simply sliding them in until they click, and the case opens with thumb-screws. Of course I showed off my build to my family (including aunt, uncle, and cousins) in addition to my best friend. People seemed most impressed with the tool-less design of the case.

I still need to purchase the third 750-GB disk to complete the RAID-5 configuration, and will probably do so when there is a good deal on NewEgg for one made by a third manufacturer. The grand total for the build comes out to just under $1200.

I would say that building my own computer from scratch was a very good experience and would recommend for anyone to do it himself for his next computer. The exception is if one is looking strictly bottom-end. These are the only computers that are a better deal if you get them from some place such as Dell or Wal-Mart.

Labels: , , , ,

Monday, January 28, 2008

The Dvorak Keyboard Layout

In the final week of Christmas vacation while I was waiting for the pieces of my server (about which I shall soon write) to be shipped to me, I happened to come across the Dvorak keyboard. This was not the first time I had heard of it -- I had previously seen mention of it in some Wikipedia user pages, but dismissed it as irrelevant at the time.

At any rate, somehow, I came across the Wikipedia Dvorak article. I found it intriguing and decided to look further into it. I imagine that the reader will have never heard of the Dvorak layout, so I will give a brief description. The Dvorak keyboard layout is an alternative to qwerty. The keys are laid out so as to minimize hand movement -- the most commonly used keys, including all of the vowels, are on the home row. Dvorak was designed for efficiency. DV Zine is a comic-style introduction to Dvorak and its history, in addition to its use.

Now, theoretically, one can type faster an a Dvorak keyboard than on a qwerty, but I'm not too shabby at typing qwerty -- I can, when concentrating on 100% accuracy, type 62 words per minute. This switch was going to cost me some time, as well as serve to make each keystroke several times more expensive -- from instant messages to Google searches to blog posts (which, granted, I don't do very often). About an hour after I started the endeavour (before I had learned more than the home row), I decided to see how fast I could do on a typing test. I scored a blazing 6 words per minute. It was a couple of days before I had nearly memorized the new layout, by which time the pieces for my server came, and I was distracted from this endeavor.

It wasn't until I came back to school that I picked up the effort again. I have since been using Dvorak exclusively, and have been taking hand-written notes in class so as to be able to keep up, and to not fall back into using qwerty. About a week ago (the last time I booted into windows, to be exact), I re-tested my speed. This time I was up near 20 words per minute. I am certainly improving, and hopefully, I'll be up to my old typing speed so that I can declare this endeavor a success.

There have certainly been some struggles in learning Dvorak. Some of these follow in no particular order. First, now that I am programming again, I frequently press the wrong keys for curly braces, the equal sign, and other such keys. Second, I often use the command-line text editor, vi, which has the entire keyboard mapped to special commands. It is annoying, for example, when I mean to save and close the document, but mistakenly delete the current line. Third, for some reason when I tell Fedora to use Dvorak by default, the volume control buttons stop working on my ThinkPad. Finally comes just the expected pains of switching layouts -- making many typos trying to use the qwerty key locations instead of the Dvorak ones.

I cannot yet fully recommend the Dvorak layout, but I will say that it's definitely worth a look. Once I have fully mastered the layout, I may at that point fully endorse it, but not until that day.

The two sites I used in my initial training were dvorak.nl and ABCD: A Basic Course in Dvorak. (And, of course, I typed this entire post with the Dvorak keyboard.)

Labels: , , , , , ,

Tuesday, October 30, 2007

MATLAB R2007b on T61 with Vista

A week or so ago, I tried to install MATLAB R2007b. The program seemed to install without a hitch, but when I went to run it, the splash screen would pop up for a fraction of a second, then it would close, and the program would not be present in memory. I tried many things to get it to work, none of which were successful. The people at the computer help desk also had no idea how to get it to work. The good news is: I just figured it out.

The solution is to install MATLAB R2007b, then uninstall both Java 6 and Java SE Development Kit 6, restart the computer, then re-install them. This will usually require re-downloading the installation files.add an environment variable called "MATLAB_RESERVE_LO" with a value of "0" (zero). After doing that, MATLAB R2007b should run perfectly well.

I hope that someone finds this useful. The computer I'm running on is the Lenovo ThinkPad T61, issued by RPI.

Update (Oct 31, 2007, 01:23 EDT): Apparently my initial solution was a fluke, and only worked by some random chance for me. After I rebooted, it no longer worked. I found this new solution from a thread where someone else was having the same problem. Apparently, it affects "Centrino" processors, which I expect means that it affects Intel's mobile processors newer than, and including, the Pentium M. I would not be surprised if the problem also affected the desktop Core 2 Duo chips as well. One would think that they would catch these things during testing before they ship, since the setup is so common.

Update 2 (Oct 31, 2007, 01:39 EDT): MathWorks, the company behind MATLAB, has an official bug report and work-around. Unfortunately, you have to create an account to see it, but the same solution (though without an explanation) is provided both here, and in the thread I referenced earlier.

Labels: , ,

Tuesday, October 23, 2007



Ever since I have been up here at Rensselaer, I have not had a proper pair of speakers. I did not have the time, nor the space in the car to bring them when my dad drove me up here in August. All music that I have listened to has been via the speakers on my laptops, or through headphones. I should mention that the speakers on my Alienware notebook are much better than average for a notebook.

When my mom came up to visit for parents weekend, she brought my set of speakers that I had left at her house. It is amazing how good they sound after only hearing notebook speakers for a few months. They aren't even very powerful speakers; they're nothing compared to the speakers in my dad's home theatre system, for instance. (I think they're like 21 watt speakers.)

Alienware Notebook

Recently, my Alienware notebook has been running very hot -- I mean to the point where I was getting stability problems. It got so bad that I couldn't even watch an entire episode of a TV show without my system crashing.

I the first thing that came to mind was that I might have applied the thermal grease incorrectly or after I replaced my processor shortly before coming up to RPI. On Friday, I decided to take it apart. Everything looked good as far as I could tell. I decided that while the computer was open, I would clean it out.

Armed with a few cotton swabs (Q-tips), some isopropyl rubbing alcohol, and some canned air (complements of my aunt and uncle), I started cleaning out the area where the fan was. I was quite surprised at the large amount of dust that came out. I started up the computer again, and low and behold: it ran much cooler.

I decided that if there were any way to test it out, it would be to run a graphic-intensive game. This was the perfect excuse for me to buy Steam's Orange Box, which includes Half-Life 2: Episode 2. I played that for about an hour at full resolution (1920x1200), and almost full settings. There were no longer any stability issues with my machine. It will probably be a long time before I get a chance to actually play and finish the game.

Because of this experience, if anyone ever has stability issues with a notebook computer, I will first recommend that he take it apart, and get all of the dust out.

Connecting to RPI's Wireless Network from Fedora

One problem that plagued me for several months was that I couldn't get my Rensselaer-issued laptop (hereafter, "ThinkPad") to connect to the wireless network here. I first tried with the GUI network tools that come with Fedora, but there was not a way to do it that way (that I could find). My next approach was to use wpa_supplicant, and connect via 802.1X PEAP-GTC. I thought this was my solution for nearly a month, but never could figure out how to do it -- a fact I attribute to the lack of documentation for configuring wpa_supplicant.

One day, I decided to post on the Facebook "Wall" of the RPI network asking if anyone was able to get it working, as I know that many people here run Linux. A few days later, I got a reply suggesting that I try the Cisco VPN client for Linux. I tried to do this, but is couldn't compile a kernel module, or some other non-sense that was way over my head. After researching this problem, I came across the mention of a program called vpnc. There were clear instructions on how to take a Cisco client configuration file, and extract the information necessary to get vpnc to connect to the VPN. At Last, I finally was able to connect to the wireless network from Fedora! The solution was to use vpnc, if I didn't make that obvious.

In other news, in my experimentation with Fedora, I got the original Half-Life to run with the proprietary nVidia driver through WINE, a Windows emulator for Linux. I was not, however, able to get any of the other games working. One day when I have some time, I may look into this further.

Desktop Computer

My desktop is once again, nearly out of disk space. In reality I have around 120 gigabytes left, but it is not continuous; it is spread over 9 disk partitions, which is not very useful for moving and saving DVD images. I have many linux images, and other large files on my desktop (which is I am basically using as a server).

As a result of this free space being fragmented across many partitions, I have been wanting to build a RAID-5 server with several terabytes of storage in one continuous blob. The main thing precluding my from doing so at this time is funding. Additionally, I don't know which controller I would get. I need one that is reliable, fast, and cheap. (I know that the saying is usually "choose any 2".) I would end up getting some large disks from NewEgg. I would also need an inexpensive case to house the whole thing.

I would probably set it up as a Linux server, and make the space available to Windows clients via the samba protocol. I have never set up such a server before, so it would be very interesting.

Labels: , , , ,

Sunday, June 24, 2007

Spell Checking in Blogger

It seems that spell checking in Blogger now works. I had complained about it previously, but am glad that they now have it fixed. Logging in and writing a post is still full of bugs in Opera, but I can manage, though I do hope that they fix it.

One thing that kind of annoyed me is that Blogger thought that "Rensselaer" was misspelled, and didn't offer any logical replacement. On the RPI message boards, there is an entire thread on misspellings of Rensselaer.

Labels: , , , ,

Resetting an iPod Nano that has Locked Up or Frozen

Microsoft software is notorious for locking up, crashing, or hanging. Today, I was going to add some music to my brother's iPod Nano when the screen froze. I tried plugging it in several times, and Windows said that the device had malfunctioned, and should be replaced.

Obviously, replacing the iPod was out of the question, as my brother had won it as a prize. I searched Google for a solution, and eventually found one around the fiftieth result. I found instructions to reset a broken iPod. The site didn't mention the Nano, but the instructions for the others did work.

Here is how I did it:

  1. Make sure the hold switch is in the off position
  2. Move the hold switch to the on position, then back off
  3. Press and hold both the "MENU" button and the center select button together for 6 to 10 seconds until you see the Apple logo
  4. Turn the iPod back on
  5. Enjoy the no-longer-hanging iPod Nano

Many people think of Apple and everything they make as perfect, but as evidenced by my experience with the iPod Nano hanging, and similar experiences of others, even Apple products are imperfect, and malfunction.

Labels: , , ,